1. This demo centers around an arXiv dataset up to August 19, 2023, which includes approximately 5500 papers exclusively dedicated to various facets of large language models.

2. The primary focus is tracking the popularity fluctuations of large language models, accomplished by pinpointing particular models mentioned in either the paper titles or abstracts.

3. The objective is to extract valuable insights into the dynamic landscape of large language models and how their prominence evolves over time.

These charts offer interactive features such as zooming in, hiding elements, hovering for more information, and resetting for a clearer understanding.

Let's consider the diffusion speed of these models. Yes, do an alignment of the above plots.

X-axes are the days after the model was proposed.